Pattanapanyasat, Kovit PhD*†; Phuang-Ngern, Yuwadee MSc†; Sukapirom, Kasama MSc*; Lerdwana, Surada BSc*; Thepthai, Charin MSc†; Tassaneetrithep, Boonrat PhD, MD*

INTRODUCTION

Since the “3 by 5” strategy was announced by the World Health Organization and the Joint United Nations Program on HIV/AIDS in December 2003,¹ more than 2.5 million HIV-infected individuals had received antiretroviral therapy (ART) by November 2007. This is up from 100,000 in December 2001.² Today, ART is becoming more accessible to people living with HIV/AIDS in developing countries particularly in east, south, and southeast Asia; this is partly due to the availability of inexpensive generic drugs.³ However, successful management of health care of persons infected with HIV in resource-limited countries is possible only if the ART program is accompanied by accessible and affordable laboratory servicing in the form of CD4⁺ T-lymphocyte monitoring.

In HIV-1-infected persons, CD4⁺ T-lymphocyte count is one of the most important indicators for assessing the degree of immune deterioration and disease progression toward AIDS; together with clinical information, the CD4⁺ T-lymphocyte count is used for defining decision to initiate ART and for monitoring the efficacy of ART for guiding regimen changes and deciding the timing for prophylaxis of opportunistic infection.^4-6 Immunophenotyping of CD4⁺ T lymphocytes by flow cytometer (FCM) is the most accepted standard technology because of its precision, accuracy, and reproducibility.^7,8 FCM can be performed using either dual-platform (DP) or single-platform (SP) method,^7-10 but the SP method is widely accepted as it introduces fewer variable factors and has the lowest intralaboratory variation when compared with the DP method.^11-14 The SP method produces the absolute CD4⁺ T-lymphocyte counts by using known numbers of fluorescent microspheres admixed to a known volume of CD4⁺ stained blood sample either by 2-color or by 3-color fluorochrome-conjugated monoclonal antibody (mAb) reagents. A good example of this bead-based FCM is FACSCount system from Becton Dickinson Biosciences (BDB). It is, at present, the oldest and the only dedicated cytometer that has been extensively validated and widely used.¹⁵ This system automatically counts the CD4⁺ and CD8⁺ T lymphocytes in a twin tube (or only CD4⁺ T lymphocytes in a single tube) containing a mixture of mAb reagents for CD4/CD3 and a known density of fluorescent microspheres. Unfortunately, the running cost of this SP bead-based method using FACSCount or other instruments using 3-color mAb reagents, combined with reference microspheres, remains relatively high and beyond the reach of most patients with HIV/AIDS in resource-limited settings. The high cost is attributable to the use of expensive mAb reagents and microspheres, essential internal quality control reagents, and costly maintenance services of the instrument. During the past decade, there have been major breakthroughs in the development of 2 affordable SP FCMs that do not require expensive mAb reagents. These are Guava Personal Cell Analyzer (PCA) from Guava Technologies (Hayward, CA) and CyFlow from Partec GmbH (Münster, Germany). Both SP FCMs provide an absolute number by counting CD4⁺ T lymphocytes in a precisely determined blood volume. There have been several studies, including our own, which showed that these 2 new volumetric technologies performed well in comparison with the performance of the standard SP bead-based FCM method.^16-19 The low cost of testing by these 2 new FCMs has already increased access to CD4 testing, particularly for HIV-1-infected persons living in developing countries or in resource-limited settings.

Enumeration of CD4 T-lymphocyte counts from HIV-infected patients' blood samples is normally monitored by 1 type of FCM. However, monitoring by different FCM systems is not uncommon particularly in the resource-limited settings as laboratories replace their old FCM with the new one. The other reason for changing FCM for monitoring CD4 T lymphocytes in HIV-infected persons is the migration of patients moving from one locality or hospital to other hospitals. Often the Guava or CyFlow (or one of other alternative technologies) is chosen as it is more affordable. For routine CD4 T-lymphocyte immunophenotyping in any clinical laboratory, changing FCM or even an upgrade of FCM and software requires validation and interlaboratory comparisons. There are presently more than 2000 sets of these 2 volumetric FCM systems in resource-limited settings.^20,21 It is likely that these 2 FCM systems are being used in countries where many existing standard BDB FACSCount or other FCMs using the standard 3-color mAb method have been in service. To our knowledge, these 2 volumetric FCM systems have never been evaluated at the same time with the standard SP bead-based and DP FCM systems. Therefore, the objective of this study is to validate the use of these 2 volumetric FCMs for determining absolute CD4 T-lymphocyte counts from HIV-infected blood samples by comparison with the standard FACSCount and the other 2 most commonly used, the SP bead-based 3-color system and the DP 3-color FCM system, to see whether these 2 volumetric FCM systems could be used interchangeably with each other or with the well-established FCM systems.

MATERIALS AND METHODS

Patients and Blood Samples

Peripheral blood samples from 150 HIV-1-seropositive patients were used in this study. Two milliliters of venous blood from each patient was collected by venipuncture into K₃EDTA-containing tubes, kept at room temperature (24-26°C) and processed for immunophenotyping within 8 hours. All HIV-infected blood samples were part of the routine clinical specimens submitted to the Department of Immunology, Faculty of Medicine Siriraj Hospital, Bangkok, Thailand, for routine lymphocyte immunophenotyping. One aliquot of EDTA blood from each sample was also analyzed for complete blood count using Coulter STKS hematology analyzer (Beckman Coulter, Miami, FL). HIV-1 infection was diagnosed at the Department of Microbiology, Faculty of Medicine Siriraj Hospital, by using serologic testing (AxSYM HIV-1/HIV-2, Abbott GmbH, Germany) with confirmation by 2 other different serologic testings (1 + 2 VITROS; Ortho-Clinical Diagnostics and Serodia HIV; Fujirebio, Inc, Tokyo, Japan). This study was approved by the Ethical Committee of the Faculty of Medicine Siriraj Hospital, Mahidol University.

Instruments Used in This Study

The FACSCount system (BDB) is a complete SP benchtop FCM system that includes instrument, reagents, controls, and the built-in software. The instrument consists of a convection-cooled green laser and employs a direct immunophenotyping method for enumerating absolute CD4⁺ and CD8⁺ T-lymphocyte counts by using 2-color mAb reagents in a twin tube containing calibrated beads and additional control beads. The first tube in each pair consists of a mixture of mAb reagents of CD4/CD3 conjugated to a phycoerythrin (PE) and PE.Cychrome (PE.Cy5) fluorescence and using a known density of fluorescent beads. The second tube contains CD8/CD3. The control set consists of fluorescent beads at 4 different densities: zero (0 beads/μL), low (50 beads/μL), medium (250 beads/μL), and high (1000 beads/μL).

The FACSCalibur system (BDB) is a multicolor benchtop FCM equipped with a dual-laser set up that includes a 15-MW air-cooled argon ion laser, which operates at 488 nm, and a 635-nm red diode laser. The blue laser is used to excite fluorescein isothiocyanate (FITC), PE, and the peridinin-chlorophyll-protein (PerCP). The second laser is for excitation of allophycocyanin fluorochrome. This system, when combined with 3 (FITC/PE/PerCP) fluorescence-conjugated TriTEST mAb reagents of CD3/CD4/CD45 and CD3/CD8/CD45 (BDB), can generate percent CD4⁺ and percent CD8⁺ T lymphocytes, which in DP approach, these percentage values are used for calculation of absolute CD4⁺ and CD8⁺ T-lymphocyte values by multiplying with the absolute lymphocyte count from hematology analyzer. For SP approach, the known density fluorescent-integrated TruCOUNT bead (BDB) tubes are used to generate absolute values of both CD4⁺ and CD8⁺ T-lymphocyte counts.

The Guava PCA is a benchtop SP machine known as a microcapillary cytometer. The system consists of a 532-nm green diode laser with a forward scatter detector and 2 fluorescence detectors of orange (580 nm) and red fluorescence (675 nm). It is used in combination with the 2-color Guava EasyCD4 (CD4-PE/CD3-PE.Cy5) and EasyCD8 (CD8-PE/CD3-PE.Cy5) mAb reagents and measurement of a volumetric control system that allows a precise count of cell numbers and measurement of fluid volume and is regulated by a variable-speed fluid (stepper motor syringe) pump that does not require sheath fluid.

The CyFlow system used in the study is a benchtop CyFlow^green FCM equipped with a 532-nm green solid-state laser featuring a PE detector. Enumeration of absolute CD4⁺ and CD8⁺ T-lymphocyte counts is based on measuring fluid volume defined by a fixed mechanical probe design and counting of immunostained lymphocytes with PE-conjugated mAb to CD4 and CD8; thus, no counting of microspheres is needed. The fixed volume (200 μL) is defined by the distance between 2 platinum electrodes in the sample tube with a given known diameter.

Immunophenotyping Staining of Peripheral Blood

All 5 immunophenotyping staining systems required the use of 2 tubes, 1 for CD4⁺ T-lymphocyte enumeration and the other tube for CD8⁺ T-lymphocyte enumeration.

For the standard FACSCount system, 50 μL of EDTA-anticoagulated whole blood was added to each of a pair of CD4/CD3 and CD8/CD3 reagent tubes by using electronic pipette. The tubes were vortexed for 5 seconds and incubated in the dark at room temperature for 60 minutes. After incubation, 50 μL of the fixative provided with the reagent kit was added to each tube and both were incubated for 5 minutes in the dark at room temperature. After vortexing the tubes, the no-lyse-stained samples were analyzed using the FACSCount FCM.

For the FACSCalibur/TriTEST SP system, 10 μL of TriTEST 3-color mAb reagents and 50 μL of EDTA-anticoagulated whole blood were added to a TruCOUNT tube containing a known microsphere concentration. The mixture was incubated for 20 minutes at room temperature in the dark before adding 450 μL of FACS Lysing Solution (BDB). After an incubation time of 15 minutes, the lyse-no-wash stained samples were analyzed by FACSCalibur FCM. The same 3-color immunophenotyping staining approach was applied in the FACSCalibur/TriTEST DP system, except that no TruCOUNT tubes were used.

For the Guava PCA system, 10 μL of EDTA-anticoagulated whole blood and 10 μL of premixed mAb solution (Guava EasyCD4 and EasyCD8 reagents consisting of 1 μL CD3, 1 μL CD4 or CD8, and 8 μL of phosphate-buffered saline) were added to 1.5-mL microfuge tubes. The mixtures were vortexed and incubated for 15 minutes at room temperature in the dark before adding 180 μL of lysing/fixing solution, making a total mixture volume of 200 μL. The mixture was incubated for an additional 15 minutes before running on the Guava PCA.

Staining with the CyFlow^green reagents was performed by mixing 100 μL of EDTA-anticoagulated whole blood and 10 μL of mAb reagents for CD4 or CD8 in Röhren polystyrene tubes (Sarstedt, Nümbrecht, Germany). The mixtures were incubated for 10 minutes at room temperature in the dark before addition of 2500 μL of no-lyse dilution buffer (Partec), making a total mixture volume of 2610 μL. From this well-mixed no-lyse-stained sample, 850 μL was transferred to another Röhren tube and analyzed by the CyFlow^green FCM.

FCM Analysis

For the FACSCount system, each stained sample was acquired according to the manufacturer's recommendation (BDB).¹⁵ Results were expressed as absolute CD3⁺, CD4⁺, and CD8⁺ T-lymphocyte values and the ratio of CD4/CD8. These values were automatically calculated from the ratio of fluorescent cells to the reference beads, multiplied by the known concentration of beads in the tube, by using the built-in software.

Both SP and DP FACSCalibur/TriTEST FCM systems were acquired and analyzed according to the BDB's MultiSET software on the FACSCalibur system. After acquiring data on 15,000 cells, a region was automatically selected when set on side scatter^low/CD45 PerCP^high+ cells. Cells in this gate were considered to be lymphocytes. Once this was established, the percentage in the DP FCM method and absolute counts in the SP FCM system of CD3⁺/CD4⁺ or CD3⁺/CD8⁺ T lymphocytes were then automatically generated by the profile obtained using CD3-FITC/CD4-PE or CD3-FITC/CD8-PE data and the MultiSET software provided.

For the Guava PCA analysis, a quadrant was selected and set on the PE.Cy5-conjugated CD3⁺ T lymphocytes and forward scatter dot plot. After acquiring at least 2000 CD3⁺ cells, a recommended number of events that would satisfy the requisite degree of statistical significance, the double positive CD4⁺/CD3⁺ or CD8⁺/CD4⁺ T lymphocytes were then automatically obtained in another quadrant plot of CD3-PE.Cy5/CD4-PE or CD3-PE.Cy5/CD8-PE. An absolute CD4⁺ or CD8⁺ T-lymphocyte count was directly obtained from calculating the number of positive cells divided by the total acquired volume and then multiplied by the dilution factor of 1:20 by using the Cytosoft software (Guava).

For the CyFlow^green analysis, the bright CD4⁺ or CD8⁺ T-lymphocyte population was manually gated by using a histogram marker set. A dilution of 26.1 obtained from the total stained sample volume of 2610/100 μL blood sample was then set in the built-in FloMAX software (Partec). The absolute CD4⁺ or CD8⁺ T-lymphocyte counts were calculated by multiplying the number of CD4⁺ or CD8⁺ T lymphocytes by the dilution factor and dividing the results by 200 as CyFlow^green counts cells in a fixed volume of 200 μL.

Quality Control and Assay Precision

In this simultaneous comparison between the 2 volumetric FCM systems and the other 3 predicate FCM systems, we endeavored to keep all the variables consistent for all FCM systems, including sample processing and acquisition and performance of instrument by using the same batch of reagents for each system throughout the study. To avoid operator-induced variations, all the immunostaining procedures and the FCM analyses were performed by the same operator for each instrument. Moreover, adequate training on the use of reverse pipetting technique and electronic pipette was also provided for each operator. In addition, the FCM photomultiplier tube voltage, sensitivity, and fluorescent compensation settings were optimized before sample acquisition and analysis using FACSCount controls, Calibrite beads (BDB), a control set of fluorochrome-integrated beads (BDB), and a Guava Check beads kit and Partec calibration beads with CountCheck for the FACSCount, FACSCalibur, Guava, and CyFlow, respectively. To assess for the precision of the 2 volumetric FCM systems, 10 fresh normal whole blood samples and 5 stabilized whole blood preparations from the CD-Check Plus CD4 Low reagent kit (Streck, Omaha, NE) were also used for within-run and between-run variations, and, when applicable, the across-instrument pooled coefficient of variations (CVs) were calculated for absolute CD4⁺ T-lymphocyte counts.

Statistical Analysis

Comparison of CD4⁺ and CD8⁺ T-lymphocyte counts obtained by different FCM systems was assessed by correlation determinations (r2) and linear regression analysis using StatView (Brainpower, Calabasas, CA). To allow the determination of whether there was any systematic bias when the absolute CD4⁺ and CD8⁺ T-lymphocyte counts derived from these 2 volumetric FCM systems were compared with those obtained by the standard FACSCount and the other 2 most commonly used SP bead-based 3-color and the DP 3-color FCM systems, the difference between each data pair of measurements (system A-system B) was graphically plotted on the vertical axis against the average of the pair (system A+ system B/2) on the horizontal axis, as described by the Bland-Altman statistical bias method.²² The mean absolute difference (bias) and the limits of agreement (LOA: equivalent to the mean difference ± 1.96 SD) were calculated. To examine the possible differences and the potential clinical impact of the absolute CD4⁺ T-lymphocyte counts obtained by these 2 volumetric FCM systems, significance tests for absolute CD4⁺ T-lymphocyte values in samples <200 cells per microliter were also determined. Percent similarity calculation between the 2 volumetric FCM systems and the other 3 predicate systems was also performed by taking the average of each method pair divided by the predicate system and multiplying by 100.²³

RESULTS

When operational precision for within-run variation of the 2 volumetric FCM systems and the 3 predicate FCM systems was determined using fresh whole blood samples and Streck CD-Check Plus CD4 Low-stabilized whole blood preparations, the mean percent CVs of CD4⁺ T-lymphocyte counts derived from running 5 replicates from fresh whole blood and 3 replicates from stabilized whole blood were <7% for each FCM system and the CyFlow^green showed a slight higher average CVs compared with other FCM systems (Table 1). For between-run reproducibility, the average CV of 3 replicates from 5 stabilized whole blood preparations, determined over the period of 1-month study, did not exceed 10%.

Table 1

The mean absolute counts of both CD4⁺ and CD8⁺ T lymphocytes for 150 HIV-1-infected blood samples analyzed by the 5 FCM systems are shown in Table 2. As can be seen, all FCM systems gave similar absolute values of CD4⁺ and CD8⁺ T lymphocytes. CyFlow^green gave a lower mean of CD4⁺ T-lymphocyte counts than other FCM systems, with the highest values obtained from the Guava system (CyFlow^green<FACSCount<DP TriTEST<SP TriTEST/TruCOUNT<Guava). For CD8⁺ T-lymphocyte counts, CyFlow^green system also showed the lowest mean followed by the Guava system (CyFlow^green<Guava<FACSCount<DP TriTEST<SP TriTEST/TruCOUNT).

Table 2

Comparisons were made between the absolute CD4⁺ T-lymphocyte values obtained from the 2 volumetric FCM systems and those from the standard FACSCount system, SP TriTEST/TruCOUNT system, and DP TriTEST system (Fig. 1). The correlation plots from the Guava system were excellent with the FACSCount system (r2 = 0.931; y = 22.113 + 1.007x, P < 0.0001; Fig. 1A), with the SP TriTEST/TruCOUNT system (r2 = 0.933; y = 20.883 + 0.954x, P < 0.0001; Fig. 1B), and with the DP TriTEST system (r2 = 0.935; y = 15.860 + 1.016x, P < 0.0001; Fig. 1C). The absolute CD4⁺ T-lymphocyte counts obtained from the CyFlow^green system were also well correlated with the 3 predicate systems, with r2 = 0.978 (y = 3.808 + 0.954x, P < 0.0001), 0.974 (y = 3.492 + 0.901x, P < 0.0001), and 0.974 (y = −0.999 + 0.959x, P < 0.0001) for the FACSCount system (Fig. 1D), the SP TriTEST/TruCOUNT (Fig. 1E), and the DP TriTEST system (Fig. 1F), respectively. As expected, correlation of the CD4⁺ T-lymphocyte counts obtained by the 2 SP bead-based systems was very high with r2 = 0.984 (y = 1.529 + 0.983x, P < 0.0001). Interestingly, the correlation between the 2 volumetric FCM systems was also high but with a lower r2 of 0.928 (y = 21.856 + 1.043x, P < 0.0001) when compared with other FCM systems.

Figure 1

Bland-Altman bias plots were analyzed by plotting the difference between the absolute CD4⁺ T-lymphocyte values generated using the Guava system and those from each of the 3 predicate FCM methods against the mean absolute values of the 2 methods (Fig. 2). The mean bias was +24.1 cells per microliter (LOA, −107.4 to +155.5 cells/μL, Fig. 2A), +6.3 cells per microliter (LOA, −125.3 to +137.9 cells/μL, Fig. 2B), and +20.5 cells per microliter (LOA, −107.2 to +148.2 cells/μL, Fig. 2C), when the Guava system was compared with the FACSCount system, the SP TriTEST/TruCOUNT system, and the DP TriTEST system, respectively. For CyFlow^green, the mean bias was toward a lower trend of −9.8 cells per microliter (LOA, −81.9 to +62.3 cells/μL), −27.6 cells per microliter (LOA, −117.2 to +62.1 cells/μL), and −13.4 cells per microliter (LOA, −89.6 to +62.9 cells/μL), when this system was compared with the FACSCount system (Fig. 2D), the SP TriTEST/TruCOUNT system (Fig. 2E), and the DP TriTEST system (Fig. 2F), respectively. These findings indicate that the volumetric Guava system yielded higher CD4⁺ T-lymphocyte counts than the 2 SP bead-based systems and the DP system. In contrast, the CyFlow^green system showed lower CD4⁺ T-lymphocyte values than the 2 SP bead-based systems and the DP system. Nevertheless, the percent similarity was high when the Guava system was compared with the FACSCount system, the SP TriTEST/TruCOUNT system, and the DP TriTEST system, with percent similarity of 114.96% (CV = 33.21), 107.42% (CV = 24.61), and 108.84% (CV = 25.76), respectively. For the CyFlow^green system, the percent similarity was also high, with similarities of 109.16% (CV = 49.07), 102.08% (CV = 42.69), and 103.32% (CV = 42.16), respectively. Of the 2 volumetric systems, the Guava system had higher CD4⁺ T-lymphocyte counts than the CyFlow^green system because the mean bias of CD4⁺ T-lymphocyte counts was +33.9 cells per microliter with LOA of −101.2 to +169.0 cells per microliter and the percent similarity was 109.05% (CV = 15.23). For the 2 SP bead-based systems, the mean bias of CD4⁺ T-lymphocyte counts was +17.8 cells per microliter with LOA of −51.5 to +87.1 cells per microliter and the percent similarity was 106.16% (CV = 13.28).

Figure 2

The bias analysis was separately performed on the 49 HIV-1-infected blood samples with absolute CD4⁺ T-lymphocyte counts ≤200 cells per microliter. Again, the CyFlow^green system showed the lowest CD4⁺ T-lymphocyte counts, and the Guava system gave the highest values (Table 3). Just as was observed in the overall data set, a bias toward higher CD4⁺ T-lymphocyte counts was observed in the Guava system with a bias of +18.4 cells per microliter (LOA, −20.1 to +57.0 cells/μL). When compared with the FACSCount system (Fig. 3A), a bias of +12.3 cells per microliter (LOA, −23.1 to +47.6 cells/μL) was found when compared with the SP TriTEST/TruCOUNT system (Fig. 3B) and a bias of +12.9 cells per microliter (LOA, −26.5 to +52.3 cells/μL) when compared with the DP TriTEST system (Fig. 3C). For CyFlow^green system, a trend toward a lower bias of CD4⁺ T-lymphocyte counts of +0.8 cells per microliter (LOA, −21.7 to +23.2 cells/μL) was observed when this system was compared with the FACSCount system (Fig. 3D). A lower bias of −5.4 cells per microliter (LOA, −28.9 cells to +18.1 cells/μL) was found when compared with the SP TriTEST/TruCOUNT system (Fig. 3E) and a bias of −4.8 cells per microliter (LOA, −27.8 to +18.2 cells/μL) when compared with the DP TriTEST system (Fig. 3F).

Table 3

Figure 3

Analysis of the linear correlation and the Bland-Altman plots of the CD8⁺ T-lymphocyte data also showed good correlation and agreement between the data from the 2 volumetric FCM systems and the other 3 predicate FCM systems with the bias toward the lower values (Table 4). The difference between the 2 standard bead-based methods was +22.0 cells per microliter (LOA, −173.7 to +217.7 cells/μL), whereas the mean bias from the 2 volumetric FCM systems was +55.6 cells per microliter (LOA, −317.3 to +428.5 cells/μL).

Table 4

DISCUSSION

Most conventional benchtop FCM systems used for enumeration of CD4⁺ T-lymphocyte counts are relatively complex, costly, and technically demanding, and the instruments used need constant maintenance. Thus, these systems are difficult to apply for routine use in most laboratories operating with poor facilities and in resource-limited settings. An ideal CD4 testing FCM for such settings would be an SP FCM approach that minimizes the sources of variations and is less expensive, is small, is relatively simple to use, and yet reproducible. Above all, it should be interchangeable with the standard FCM system. The 2 affordable volumetric Guava PCA and CyFlow^green FCM systems potentially satisfy most of these criteria and were the focus of this study. They were evaluated for their performance and compared with the standard SP bead-based FACSCount system and the TriTEST/TruCOUNT system and the DP TriTEST system because these are the SP and DP systems that have long been widely used in many countries, including resource-limited settings such as Thailand.

In this study, absolute CD4⁺ and CD8⁺ T-lymphocyte values obtained from the 2 volumetric FCM systems correlated highly (R2 ≥ 0.928) with those from standard bead-based 2-color FACSCount system and the TriTEST/TruCOUNT using FACSCalibur and the DP TriTEST using FACSCalibur FCMs. The overall biases for absolute CD4⁺ T-lymphocyte values were about +25 cells per microliter more for the Guava system and −30 cells per microliter less for the CyFlow^green system. This indicates that the absolute CD4⁺ T-lymphocyte values from the Guava system are overestimated, whereas the absolute values obtained from the CyFlow^green system are underestimated. If one considers HIV-infected individuals who need ART because their absolute CD4⁺ T-lymphocyte counts are less than 200 cells per microliter, the CD4⁺ T-lymphocyte count biases are 18.4, 12.3, and 12.9 cells per microliter higher with the Guava system than the FACSCount system, the TriTEST/TruCOUNT system, and the DP TriTEST system, respectively. These will result in CD4⁺ T-lymphocyte count biases that are about 6%-9% higher when the Guava system is used, and although these high-bias figures might influence the clinical decision making for HIV-1-infected individuals, they are unlikely to affect monitoring as long as the systems are not used interchangeably. However, if a replacement of the existing standard FCM systems with the Guava system is unavoidable, it is recommended that the laboratories should consider the impact for such variations and, when possible, have a result based on the standard FCM system and the Guava system placed in parallel on at least 1 sample for each patient, particularly if the patient's absolute CD4⁺ T-lymphocyte count may be in the vicinity of a crucial therapeutic decision point. On the contrary, the CyFlow^green system showed the CD4⁺ T-lymphocyte count biases of only 0.8, 5.4, and 4.8 cells per microliter, less than the FACSCount system, the TriTEST/TruCOUNT system, and the DP TriTEST system, respectively. Overall, these biases are less than 3% when the CyFlow^green system is used. Though these differences may not seem to have any clinical significance, it is suggested that the testing laboratories should each evaluate their own instrument performance to determine if the new instrument could be interchanged with the old and the existing standard FCM system. Because the performances of the Guava and the CyFlow^green systems are substantially different, it is important, therefore, not to replace either of these 2 volumetric FCM systems with another volumetric FCM system.

Although data from the 2 volumetric FCM systems showed good correlation with the predicate FCM systems throughout the whole study range, there were some consistent biases, with higher absolute CD4⁺ T-lymphocyte values for the Guava system and lower absolute CD4⁺ T-lymphocyte values for the CyFlow^green system. Our study is in line with previous evaluation studies of higher values when using the Guava system^19,24,25 and lower values for the CyFlow^green or other CyFlow system.^17,18,26,27 These variations may be due to technical differences in the systems. All SP FCM systems require a high level of precision in the dispensing of reagents and at all the pipetting steps, with particular care regarding dilution steps,^13,28 so these variations in pipetting steps might attribute to their differences. In the 2 standard SP bead-based systems, the microspheres used are preloaded and strictly prepared and controlled by the manufacturer; there should not be any biases in relation to the reference microspheres. The Guava system, on the other hand, defines its absolute CD4⁺ T-lymphocyte counts in a preset final volume of cell suspension. Hence, calculations may be affected by any factor influencing concentrations. For CyFlow^green system, factors that may influence the values obtained are the use of different red blood cell lysing or fixing reagents in the assay systems. It is known that lysing reagents may affect leukocyte populations differently; some reagents may cause significant loss of leukocytes that may contribute to error, especially in samples from HIV-infected patients whose cells may be more susceptible to lysis.^13,29 Artifacts such as cell debris, cell aggregate, background fluorescence, and, more importantly, the interference by some CD4⁺ monocytes contaminating in the lymphocyte boundary may further lead to lymphocyte impurities and thus may influence the CD4⁺ T-lymphocyte count accuracy.

This study showed that the 2 volumetric FCM systems are valid for CD4⁺ T-lymphocyte enumeration. Nevertheless, because this evaluation study was a single laboratory's experience, and performed in a reference laboratory with the resources and technical skills that fulfill all required FCM standards, a more complete comparison, involving intralaboratory and interlaboratory validation, should be conducted to determine if such 2 volumetric systems can be interchanged with other existing systems of CD4⁺ T-lymphocyte counting used in resource-limited settings. Although the evaluation suggests that the 2 systems are promising compared with the existing standard bead-based FCM systems, there are some limitations of the systems and challenges to their implementation. Operation of these 2 systems still requires substantial technical expertise and good maintenance. For example, calibration of the CyFlow^green instrument requires tedious and manual adjustment of calibration beads, whereas the easily breakable microcapillary flow cell in the Guava system needs frequent maintenance because the absence of a need for sheath fluid in the Guava system means that the system's sampling precision depends solely on the integrity of the fluid pathway. In addition, the small blood volume used in the Guava microcapillary FCM system may be problematic in samples from patients with leukopenia or patients with very low levels of CD3⁺ T-lymphocyte counts because acquiring the required 2000 CD3⁺ T lymphocytes in 4 minutes may be challenging and may generate statistical errors in determining accurate CD4⁺ or CD8⁺ T-lymphocyte counts. Although volume is not a problem in the CyFlow^green system, the limited sample volume of 200 μL and the high acquiring rate may well result in underestimation of the CD4⁺ T-lymphocyte values in samples with lymphopenia and very low CD4⁺ T-lymphocyte counts. Moreover, setting gates to distinguish both CD3⁺ and CD4⁺/CD3⁺ or CD8⁺/CD3⁺ T lymphocytes from non-CD3⁺ and non-CD4⁺/CD3⁺ or CD8⁺/CD3⁺ T lymphocytes can be difficult to establish because the intensities of positive CD4-PE.Cy5/CD3-PE or CD8-PE/CD3⁺-PE staining in the Guava system tend to be lower than that observed with the FACSCount and TriTEST/TruCOUNT systems. Although this dim fluorescence of the mAb reagents is not the problem in the case of the CyFlow^green system, the overlapping of both CD4⁺ T lymphocytes and monocytes expressing CD4 antigen histograms makes it difficult to differentiate the pure CD4⁺ T-lymphocyte population within this single-parameter CyFlow^green system.^13,30 It is suggested that blood samples with significantly high monocyte counts should be tested by the CyFlow^green system.

With respect to sample throughput and cost, each of these 2 systems offers a high throughput of 100 samples per day that are compatible with the standard FACSCount system. And their compact size means that they can easily be moved between laboratories, which are important in some localities where laboratories may have to share the facilities. In addition, the costs of mAb reagents used in these 2 systems are relatively cheap, which can reduce the cost of CD4 testing from $15-$20 per test for the standard bead-based FACSCount or TriTEST/TruCOUNT FCM systems to $3-$5 per test. The price quoted refers to the cost per reportable result, which includes all consumables, staff salary, and overhead, such as electricity.

In conclusion, these 2 volumetric FCMs are promising systems for performing SP absolute CD4⁺ T-lymphocyte counts with excellent reproducibility and correlated well with the standard SP bead-based and DP FCM systems. These cost-saving volumetric systems should facilitate wider access to CD4 testing for persons with HIV infection in resource-limited countries. However, variations among the 5 systems should also be considered if either one of these systems is interchanged with another system.

REFERENCES